The muscular dystrophies (MD) are a heterogeneous group of inherited disorders characterized by progressive weakness and degeneration of skeletal muscles. The molecular basis of Duchenne muscular dystrophy (DMD) was first elucidated twenty years ago as a perturbation of dystrophin (Koenig et al., (1987) Cell 50: 509-517). Dystrophin associates with a number of proteins to form a large oligomeric complex named the dystrophin-glycoprotein complex (DGC), which bridges across the sarcolemma and connects the extracellular matrix and the actin cytoskeleton (Allamand and Campbell, (2000) Human Molecular Genetics 9:2459-2467). Loss or abnormal function of the DGC components will lead to dystrophic muscles and various forms of MD. For instance, mutations in sarcoglycan (SG) genes are responsible for autosomal recessive limb-girdle muscular dystrophies (LGMD 2C-2F; reviewed by Lim and Campbell, (1998) Neurology 11:443-452), and deficiency in the laminin α2 chain is responsible for about half of the cases of congenital muscular dystrophy (CMD; Helbling-Leclerc et al., (1995) Nature Genetics 11:216-218).
Adeno-associated virus (AAV) was first reported to efficiently transduce muscle over ten years ago (Xiao et al., (1996) J. Virology 70:8098-8108). As another advantage, AAV vectors have a good safety profile. The recombinant AAV (rAAV) genome composed of a foreign expression cassette and AAV inverted terminal repeat (ITR) sequences exists in eukaryotic cells in an episomal form that is responsible for persistent transgene expression (Schnepp et al., (2003) J. Virology 77:3495-3504). No human disease has been associated with wild-type AAV infection and low toxicity is observed in human subjects following muscle transduction by rAAV (Manno et al., (2003) Blood 101:2963-2972).
A series of new AAV serotypes have been identified from humans or primates that display variable capsid sequences as compared with AAV2 (Gao et al., (2002) Proc. Nat. Acad Sci. USA 99:11854-11859; Gao et al., (2004) J. Virology 78:6381-6388). Of these, AAV1, 6, 8, and 9 recombinant vectors have been reported to result in higher transgene expression level in muscle than rAAV2 vector (Wang et al., (2005) Nature Biotech. 23:321-328; Inagaki et al., (2006) Molecular Therapy 14:45-53). Widespread transduction of cardiac and skeletal muscle has been achieved in adult mouse by intravenous administration of rAAV6 vector supplemented with vascular endothelium growth factor (VEGF) (Gregorevic et al., (2004) Nature Med. 10:828-834). rAAV1 vectors have a similar capsid sequence and were successfully applied in systemic gene delivery to vital muscles of DMD and CMD mouse models and efficiently ameliorated the dystrophic phenotype (Qiao et al., (2005) Proc. Nat. Acad. Sci. USA 102:11999-12004; Denti et al., (2006) Proc. Nat. Acad. Sci. USA 103:3758-3763). Whilst vector administration in these studies was conducted with pharmacological interventions or in neonatal animals, rAAV8 vectors appear more efficient at crossing the blood vessel barrier and transducing heart and skeletal muscle of adult mice and hamsters (Wang et al., (2005) Nature Biotech. 23:321-328; Zhu et al., (2005) Circulation 112:2650-2659). AAV9 vectors also demonstrate efficient tropism to the myocardium and serve as another alternative for systemic gene delivery to heart (Inagaki et al., (2006) Molecular Therapy 14:45-53; Pacak et al., (2006) Circulation Research 99:3-9).
In addition to the investigation of natural AAV serotypes, research efforts have explored modification of the AAV capsid to produce optimized vectors. Mutagenesis represented the initial approach to genetically modify the AAV2 capsid (Wu et al., (2000) J. Virology 74:8635-8647; Lochrie et al., (2006) J. Virology 80:821-834). Insertion of peptides from phage display libraries into the AAV capsid protein proved to be another strategy to modify the AAV capsid and retarget the vector to new cells or tissues. This method has been further developed to directly display synthesized peptides on the surface of the AAV capsid.
Besides rational design, directed evolution has been used to introduce modifications into AAV vectors. One group reported that the AAV2 capsid gene was diversified by random mutagenesis and then subject to in vitro recombination (Maheshri et al., (2006) Nature Biotech. 24:198-204). The modified capsid genes were employed for the production of an AAV library, which was screened in vitro for enhanced properties such as altered affinities for heparin or evasion of antibody neutralization. In vitro screening methods, however, are inherently limited in their ability to identify optimized mutants that will have desired properties in vivo in the context of a complex biological system. For example, the vasculature is a major barrier for systemic AAV delivery via the circulation to many tissues and cell types including skeletal muscle, diaphragm muscle, the heart and brain. It would be desirable for an AAV vector to not only be efficient in crossing the endothelial lining to reach the intended target cells such as cardiomyocytes in the heart, myofibers in skeletal muscle and neurons in the brain, but also be robust in infecting those intended cells after reaching them. An in vitro panning system is simply unable to select for both of these properties. In addition, there are numerous examples in the literature in which in vitro assessment of viral properties such as tropism was not predictive of in vivo behavior.